Fireworks AI

Run and customize open-source AI models with top speed and efficiency. Fireworks’ SoTA training and inference take you beyond the frontier, transforming open models into your specialized intelligence. Fireworks AI supports over 100 open-source models like Llama 3.1, DeepSeek R1, and Stable Diffusion XL across text, image, and multimodal formats. It offers a pay-per-token model with free credits for new users, competitive with industry standards. FireOptimizer supports LoRA fine-tuning for customizing models at no extra cost. It offers 99.99% API uptime and scales for high-throughput apps. Prompt caching reduces time-to-first-token by 5-10x for long, repeated prompts. It supports text, image, audio, and multimodal models for diverse applications.